Ming-Hsuan-Tu

Does this project support multi-gpu?

How to reach 22 fps in jetson nano

1

I can't reproduce t 22 FPS in jetson nano. I got only 16 FPS with resnet18 backbone. any idea?

How to tune the loss weights?

30

Hi, there are some of the loss weights, like https://github.com/yzd-v/FGD/blob/master/mmdet/distillation/losses/fgd.py#L27 How did you tune these loss weights? Any guideline to tune these hyper parameters?

Can we futher improve autoalim without gate?

3

It is not easy to deploy gate operator with some other backends, like TensorRT. So my question is can we futher improve autoalim without the dynamic gate when inference?Any ongoing...

This would fix cuda memory access error when reading the images from the disk. for example, if the desired resized resolution is 300x300, and the image being read is 1080...

Have you ever tried to train with cifa10

6

I tried to train with cifar10 and set margin to 2, but got nan. learning rate is set to 0.1. Any idea?

memory leak in boxlist_iou?

1

I found out that memory leak may happen in `box_list_iou` (https://github.com/zhangxiaosong18/FreeAnchor/blob/master/maskrcnn_benchmark/modeling/rpn/free_anchor_loss.py#L108) when it processes large number of ground truth, of course the gpu memory grows up. but it never frees...

Hi, Since FPN has a equation to map rois to different levels. larger rois would be fed into deeper layers. In the code (https://github.com/unsky/FPN/blob/master/lib/rpn/proposal_target_layer.py#L164), If there are no rois can...

the meaning of depth div

1

Hi what is the meaning of `depth_div` used in `new_ch`? https://github.com/zsef123/EfficientNets-PyTorch/blob/master/models/effnet.py#L81

Can't make the doc html ``` [autosummary] generating autosummary for: cmake_example.rst, index.rst Failed to import 'cmake_example': no module named cmake_example building [mo]: targets for 0 po files that are out...